Lane Line Segmentation using Watershed Algorithm

Abstract

Transportation is getting more expensive due to the increase in the vehicle population. On a day-to-day life, the automobile field is getting more intelligent and robust. Lane lines represent the rules for mobile vehicles. Identification of these lane lines is a challenging and advanced concept in the field of computer vision. This work may be applicable for auto-drive engine vehicles. This paper represents the road lane line segmentation using a watershed segmentation algorithm. The algorithm is tuned mainly for highways. The readily capture colored images are treated as input images. Then watershed segmentation algorithm is fed with the morphology of the image instead of markers and colored to differentiate from the other part of the image. This work is irrespective of input image size. We evaluated the accuracy of segmentation at 99%. Can enhance the same result to the video stream.